A method and system for recommending a pick-up point for carpooling in an intercity network

Abstract

The application provides a carpooling pickup point recommendation method and system for an intercity network, comprising: acquiring intercity traffic network data; projecting a three-dimensional road network to a two-dimensional topological plane; calculating a dynamic weight of each road based on real-time traffic data; folding secondary roads according to the dynamic weight and a folding threshold to generate a backbone network; generating a candidate recommendation result on the backbone network, including a mandatory point, a selectable point and a dynamic adjustment point; and optimizing and screening the candidate recommendation result to obtain a recommended pickup point. Through dynamic weight calculation and backbone network generation, the application can reflect the changes in traffic conditions in real time, generate more accurate recommended pickup points, improve the efficiency of carpooling, meet the carpooling needs of different passengers through the generation of diversified candidate recommendation points and the optimization of candidate results, improve the satisfaction of passengers, and effectively improve the calculation efficiency through the generation of the backbone network and the optimization of the candidate points, so that the recommendation system can respond to passenger requests more quickly.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of online carpooling, and in particular to a method and system for recommending pickup points for intercity online carpooling. BACKGROUND

[0002] With the acceleration of urbanization, intercity travel demand is increasing, and online carpooling, as a convenient and flexible mode of transportation, has been favored by the majority of passengers. However, in the context of intercity online carpooling, the choice of pickup point is crucial to the efficiency of carpooling and passenger satisfaction.

[0003] Traditional pickup point recommendation methods are often based on static geographic information or simple rules, which are difficult to reflect changes in traffic conditions in real time, resulting in inaccurate recommended pickup points and affecting the efficiency of carpooling and passenger experience. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a method and system for recommending pickup points for intercity online carpooling to solve the above-mentioned problems.

[0005] According to the method for recommending pickup points for intercity online carpooling, the method comprises:

[0006] Obtaining intercity traffic network data, including road nodes, edges and weight information;

[0007] Based on real-time traffic data, calculating the dynamic weight of each road;

[0008] Folding the secondary roads according to the dynamic weight and the folding threshold to generate a backbone network;

[0009] Generating a candidate recommendation result containing candidate points on the backbone network, including mandatory points, optional points and dynamically adjusted points;

[0010] Optimizing and screening the candidate recommendation result to obtain the final recommended pickup point.

[0011] Further, the step of obtaining intercity traffic network data comprises:

[0012] Obtaining GIS road data of the target area, including node coordinates, road length and road grade;

[0013] Establishing an initial traffic network graph G=(V,E), where V is a set of intersection nodes and E is a set of road edges;

[0014] Assigning an initial weight w(e) to each edge e E 初始 = road grade x basic speed coefficient.

[0015] Furthermore, the step of calculating the dynamic weight of each road includes:

[0016] Connect to the real-time traffic data interface to obtain the current average vehicle speed v of each road segment. real (e);

[0017] The dynamic weights are calculated using the following formula:

[0018] w(e) = α‧v real (e)‧δ(t)+β‧Road grade‧θ(e)+γ‧Accident probability coefficient‧φ(w),

[0019] Where δ(t) is the time penalty coefficient, θ(e) is the road function coefficient, and φ(w) is the weather impact coefficient.

[0020] Furthermore, the step of generating the backbone network by folding secondary paths according to dynamic weights and folding thresholds includes:

[0021] Set the initial folding threshold;

[0022] Traverse all edges e∈E;

[0023] The dynamic calculation of the folding threshold θ = average dynamic weight × first adjustment coefficient;

[0024] If the dynamic weight w(e) < the folding threshold θ, then a folding operation is performed to fold the nodes v at both ends of edge e. i and v j Merge into supernode V new ;

[0025] Update the set of associated edges to include the original edge set connected to node v. i and v j Connected edges are redirected to supernode V new And recalculate the dynamic weights of the affected edges;

[0026] Repeat the iteration until the network size converges.

[0027] Furthermore, the nodes v at both ends of edge e... i and v j Merge into supernode V new The steps include:

[0028] Calculate the V of the supernode new The latitude coordinates are given by the formula:

[0029] ,

[0030] Where, φ new m is the dimension of the supernode, and m is the dimension of node v. i The number of associated roads, φ i For node vi latitude coordinates, f i，a For node v i The traffic flow weight of the a-th associated road, where n is the value of node v. j The number of associated roads, φ j For node v j latitude coordinates, f j，b For node v j The traffic flow weight of the associated road in Article b;

[0031] Calculate the V of the supernode new The formula for longitude coordinates is:

[0032] ,

[0033] Where, λ new Let λ be the longitude of the supernode. i For node v i longitude coordinates, λ j For node v j Longitude coordinates;

[0034] node v i and v j All properties are inherited by V new This includes the connected roadside, traffic signal information, and historical data;

[0035] The original connection node v i and v j Edge e is removed from the network graph.

[0036] Furthermore, the step of generating candidate recommendation results containing each candidate point on the backbone network, wherein the candidate points include mandatory points, optional points, and dynamically adjusted points, includes:

[0037] Select all nodes with a degree greater than or equal to the degree threshold as core boarding points and designate them as mandatory points;

[0038] For each pair of mandatory points, the midpoint of the shortest path is selected as a candidate boarding point and thus defined as an optional point.

[0039] Based on the real-time folding status, temporary recommended points are generated within the neighborhood of the supernode as dynamic adjustment points.

[0040] Furthermore, the step of generating temporary recommended points within the supernode's neighborhood based on the real-time folding state, as dynamic adjustment points, includes:

[0041] With supernode V new An initial circular neighborhood is established centered on the node V_new. Based on the Voronoi diagram constraints, the service range of the adjacent supernode V_new is excluded.

[0042] According to the number of current unmatched carpool requests and the number of available vehicles, the neighborhood radius is dynamically adjusted, and the formula is:

[0043] ,

[0044] where r is the dynamically adjusted neighborhood radius, r0 is the initially set reference neighborhood radius, k is the second adjustment coefficient, N dynamic is the number of current unmatched carpool requests, and N requests is the number of available vehicles. vehicles

[0045] Based on the neighborhood, temporary recommended points are generated with multiple objectives;

[0046] The priority of each temporary recommended point is calculated based on traffic matching degree, demand density and facility importance;

[0047] The candidate points are dynamically adjusted.

[0048] Further, the step of generating temporary recommended points based on the neighborhood with multiple objectives comprises:

[0049] Sampling the flow gradient along the neighborhood boundary road, and selecting the position with an absolute value of flow gradient exceeding the gradient threshold as a temporary recommended point;

[0050] Applying a clustering algorithm to cluster the passenger request coordinates, and selecting a clustering core point as a temporary recommended point;

[0051] Identifying POIs, and selecting a POI close to the super node V new as a temporary recommended point.

[0052] Further, the step of optimizing and screening the candidate recommended results comprises:

[0053] Establishing a candidate pickup point set S={mandatory points+optional points+dynamically adjusted points};

[0054] For each carpool request, the walking distance cost C walk , vehicle detour cost C detour and passenger matching degree P match to each point in the candidate pickup point set S are calculated.

[0055] The comprehensive cost C is calculated by a comprehensive cost function, and the point with the minimum comprehensive cost C is selected as the recommended pickup point, and the formula is:

[0056] C=ω1C walk +ω2C detour -ω3P match , where ω1, ω2 and ω3 are the weights of walking distance cost, vehicle detour cost and passenger matching probability, respectively. ​

[0057] The application further provides a pickup point recommendation system for carpooling of intercity network ride-hailing, which is used for implementing the pickup point recommendation method for carpooling of intercity network ride-hailing.

[0058] The data acquisition module is used for acquiring intercity traffic network data, including road node, edge and weight information.

[0059] The dynamic weight module is used for calculating the dynamic weight of each road based on real-time traffic data.

[0060] The folding module is used for folding the secondary roads according to the dynamic weight and the folding threshold to generate a backbone network.

[0061] The candidate generation module is used for generating a candidate recommendation result containing candidate points on the backbone network, the candidate points including mandatory points, optional points and dynamic adjustment points.

[0062] The recommended point generation module is used for optimizing and screening the candidate recommendation result to obtain the final recommended pickup point.

[0063] In summary, the pickup point recommendation method for carpooling of intercity network ride-hailing acquires intercity traffic network data, including road node, edge and weight information, which is used for representing the structure and connection relationship of the road network; calculates the dynamic weight of each road based on real-time traffic data, which can reflect the change of traffic conditions in real time and improve the timeliness and accuracy of the recommended pickup point; folds the secondary roads according to the dynamic weight and the folding threshold to generate a backbone network, which simplifies the network structure, improves the calculation efficiency and makes the selection of the recommended pickup point focus on the key positions, thereby improving the accuracy of the recommendation; generates a candidate recommendation result containing candidate points on the backbone network, the candidate points including mandatory points, optional points and dynamic adjustment points, so as to comprehensively consider the network structure, traffic conditions and passenger demand, generate diversified recommended points, meet the carpooling demand of different passengers, improve the flexibility and convenience of carpooling, and meanwhile, the introduction of the dynamic adjustment points makes the recommended result adapt to the change of traffic conditions in real time; optimizes and screens the candidate recommendation result to obtain the final recommended pickup point, which can comprehensively evaluate and screen the candidate points by considering the walking distance cost, vehicle detour cost and passenger matching probability, and finally obtain the recommended pickup point which can not only consider the convenience of passengers but also take into account the driving efficiency of vehicles and the matching probability of passengers, thereby improving the overall efficiency of carpooling and passenger satisfaction.

[0064] The application can reflect the change of traffic conditions in real time, generate more accurate recommended drop-off points, improve the efficiency of carpooling, meet the carpooling needs of different passengers and improve the satisfaction of passengers by generating diversified candidate recommended points and optimizing the candidate recommended results, and the generation of the backbone network and the optimization of the candidate points effectively improve the calculation efficiency, so that the recommendation system can respond to passenger requests more quickly.

[0065] Additional aspects and advantages of the application will be set forth in part in the description which follows, and in part will become apparent to those skilled in the art upon examination of the following and / or can be learned by practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0066] The above and / or additional aspects and advantages of the application will become apparent and be readily understood by referring to the detailed description, taken in conjunction with the accompanying drawings, in which:

[0067] Figure 1 A flowchart of a carpool drop-off point recommendation method for intercity network car-hailing according to an embodiment of the application;

[0068] Figure 2 A system block diagram of a carpool drop-off point recommendation system for intercity network car-hailing according to an embodiment of the application. DETAILED DESCRIPTION

[0069] In order to facilitate the understanding of the application, the application will be described in detail below with reference to the relevant drawings. The drawings show several embodiments of the application. However, the application can be realized in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the application more thorough and comprehensive.

[0070] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there can be a middle element. When an element is referred to as being "connected" to another element, it can be directly connected to the other element or there can be a middle element. The terms "vertical", "horizontal", "left", "right", and similar expressions used herein are for illustrative purposes only.

[0071] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the application belongs. The terminology used in the specification of the application herein is only for the purpose of describing specific embodiments and is not intended to limit the application. The term "and / or" used herein includes any and all combinations of one or more of the associated listed items.

[0072] Embodiment one

[0073] Please refer toFigure 1 The application provides a carpool pickup point recommendation method for an intercity network car service.

[0074] S101, acquiring intercity traffic network data, including road node, edge and weight information.

[0075] It should be noted that the core data of the intercity traffic network includes road node, edge and weight information, etc. These data are the key data basis for subsequent steps of graph theory modeling, dynamic weight calculation, backbone network generation and candidate recommendation result optimization.

[0076] The node data contains basic attributes (such as latitude and longitude coordinates), topological characteristics (including node connectivity, number of adjacent edges, etc., and the node connectivity of the intersection is greater than that of the ordinary node), and extended attributes (whether it contains a transportation hub such as a high-speed rail station or an airport, whether it has a multi-modal transfer function, etc.).

[0077] The edge data contains physical attributes (such as road length, number of lanes, physical isolation type, etc.), hierarchical system (which can be classified according to the Technical Standards for Highway Engineering, expressway > first-class highway > second-class highway > third-class highway > fourth-class highway) and traffic parameters (such as design speed, speed limit value, historical average daily traffic volume, etc.).

[0078] The weight information contains an initial weight equal to the road grade x the basic speed coefficient, and a dynamic weight interface is also set in the weight information layer, including real-time speed, accident probability, weather influence and other dynamic parameter access ports. The weight is refreshed every preset time, such as 5 minutes / time.

[0079] Further optionally, the step of acquiring intercity traffic network data comprises:

[0080] Acquiring GIS road data of the target area, including node coordinates, road length and road grade;

[0081] Establishing an initial traffic network graph G=(V, E), wherein V is a set of intersection nodes, and E is a set of road edges;

[0082] Assigning an initial weight w(e) to each edge e E. 初始 = road grade x basic speed coefficient.

[0083] Understandably, the graph theory model G=(V, E) is established by GIS road data, which converts the physical road system into a calculable mathematical structure. The node coordinate precision needs to reach sub-meter level (typical value: ±0.5m), to ensure the accuracy of the network topology. The road grade classification needs to comply with the Technical Standards for Highway Engineering (such as expressway > first-class highway > second-class highway).

[0084] Establish a static attribute baseline and assign an initial weight to each edge e∈E, using the formula: w(e) 初始 = Road grade × Basic speed coefficient, realizing a two-factor coupled representation, where the basic speed coefficient needs to be obtained through traffic surveys, for example, 85% of the design speed is taken as the benchmark for expressways.

[0085] The road network is transformed into a directed or undirected graph G=(V,E) to form a static network map, providing a benchmark for subsequent dynamic adjustments.

[0086] S102 calculates the dynamic weight of each road based on real-time traffic data.

[0087] It should be noted that by considering real-time traffic data (such as vehicle speed, accident probability, etc.) and static road attributes (such as road grade), the dynamic weight of each road is calculated. The dynamic weight calculation can reflect changes in traffic conditions in real time, improving the timeliness and accuracy of recommended pick-up points.

[0088] Further optionally, the step of calculating the dynamic weight of each road includes:

[0089] Connect to the real-time traffic data interface to obtain the current average vehicle speed v of each road segment. real (e);

[0090] The dynamic weights are calculated using the following formula:

[0091] w(e) = α‧v real (e)‧δ(t)+β‧Road grade‧θ(e)+γ‧Accident probability coefficient‧φ(w),

[0092] Where δ(t) is the time penalty coefficient, θ(e) is the road function coefficient, and φ(w) is the weather impact coefficient.

[0093] Understandably, when calculating the dynamic weight formula, through v real (e) The interface implements second-level data updates, for example, with an update frequency of 10-30 seconds. The time penalty coefficient δ(t) needs to be combined with traffic flow theory, for example, using the BPR function: δ(t)=1+0.15*(V / C). 4 The road function coefficient θ(e) needs to reflect the priority of road planning, for example, θ=1.5 for main roads and θ=0.6 for branch roads. The weather impact coefficient φ(w) needs to be linked with meteorological data, such as φ=1.5-2.0 for rainy and snowy weather.

[0094] The α parameter can represent the real-time vehicle speed v real (e) It is associated with road grade to realize non-linear mapping of vehicle speed-weight, that is, the fusion of dynamic and static attributes. When α=1, the real-time vehicle speed directly affects the weight proportionally; when α>1, the sensitivity of vehicle speed is enhanced.

[0095] The road level is nonlinearly amplified by the θ(e) function (e.g., θ=1.5 for the main road and θ=0.6 for the branch road), and the β parameter plays a role in strengthening the road level, maintains the basic role of the road level, and enhances the advantage of high-level roads in dynamic calculation.

[0096] The γ parameter, as a dynamic adjustment factor, reflects the inhibitory effect of accident risk on path selection, and can couple the accident probability with the weather influence φ(w) to automatically amplify the influence of the accident probability in rainy and snowy weather (φ>1). According to the road safety rating, real-time traffic events, and insurance actuarial data, the γ value is adjusted, such as increasing the γ value for accident black spot road sections, dynamically increasing the γ value for sudden accidents, and increasing the γ value for high-risk road sections.

[0097] The calculated dynamic weight integrates three dimensions of physical properties (i.e., road level), real-time state (i.e., vehicle speed), and environmental variables (i.e., weather), and forms a double-layer network model of static reference and dynamic disturbance through calculation of dynamic weight, which can improve the fidelity of representation, so that the system can not only maintain the inherent properties of the road network, but also respond to the complex changes of the traffic system in real time, providing an accurate dynamic network model for subsequent network folding and pick-up point recommendation.

[0098] S103, folding the secondary road according to the dynamic weight and the folding threshold to generate a backbone network.

[0099] It should be noted that the secondary road is folded according to the dynamic weight and the folding threshold to generate a backbone network, so as to remove redundant information in the network, retain key road nodes and connection relationships, and simplify the network structure. The generation of the backbone network not only improves the calculation efficiency, but also makes the selection of the recommended pick-up point more focused on the key positions, thereby improving the accuracy of the recommendation.

[0100] Further, the step of folding the secondary road according to the dynamic weight and the folding threshold to generate a backbone network comprises:

[0101] setting an initial folding threshold;

[0102] traversing all edges e∈E;

[0103] dynamically calculating the folding threshold θ=average dynamic weight×first adjustment coefficient;

[0104] if the dynamic weight w(e) is less than the folding threshold θ, performing a folding operation to merge the nodes v i and v j at both ends of the edge e into a super node V new ;

[0105] updating the associated edge set to replace the original edges connected to the nodes v i and v jThe connected edges are redirected to the supernode V new and the dynamic weight of the affected edges is recalculated.

[0106] The iteration is repeated until the network size converges.

[0107] It is understood that the initial folding threshold is first set as a benchmark for judging the importance of the road; all edges e E in the network are traversed, the dynamic weight w(e) of each edge is calculated, which is based on three-dimensional dynamic factors of physical attributes (i.e. road level), real-time state (i.e. vehicle speed), and environmental variables (i.e. weather); and the folding threshold θ is dynamically calculated based on the average dynamic weight and the first adjustment coefficient, which dynamically adjusts with the network state to ensure the adaptability of the folding process; if the dynamic weight w(e) of an edge e is less than the folding threshold θ, the folding operation is performed: the two end nodes v i and v j of the edge e are merged into a supernode V new , the folding operation essentially removes the secondary road from the network and simplifies the network structure by merging nodes; the associated edge set is updated, and all edges originally connected to v i and v j are redirected to the supernode V new to ensure that the connectivity and topological relationship of the network remain after folding, and the dynamic weight of the affected edges is recalculated; the above iteration process is repeated until the network size converges, i.e. the number of nodes is stable, indicating that there are no more secondary roads to fold, and the backbone network is generated. The folding based on dynamic weight and threshold can respond to changes in the traffic network in real time, ensuring the timeliness and accuracy of the backbone network. By merging secondary roads, retaining key nodes and edges, and generating a concise backbone network, efficient implementation of traffic planning, path navigation and other applications is facilitated. The iterative folding process gradually simplifies the network, reduces computational complexity, and improves algorithm running efficiency.

[0108] Further optionally, the step of merging the two end nodes v i and v j of the edge e into a supernode V new includes:

[0109] The latitude coordinate of the supernode V new is calculated, and the formula is:

[0110] ,

[0111] where φ new is the latitude of the supernode, m is the number of roads associated with the node v i , φ i is the latitude coordinate of the node v i , and f i，a is the latitude coordinate of the node v ithe traffic weight of the a-th associated road of node v, n is the number of nodes v j the number of associated roads, φ j the latitude coordinate of node v j j，b the traffic weight of the b-th associated road of node v j

[0112] calculating the V new of the super node, the formula is:

[0113] ,

[0114] where λ new is the longitude of the super node, λ i is the latitude coordinate of node v i , λ j is the latitude coordinate of node v j ;

[0115] inheriting all the attributes of nodes v i and v j to the V new , including the connected road edges, traffic signal information and historical data;

[0116] removing the edge e originally connecting nodes v i and v j from the network graph.

[0117] It can be understood that the latitude of the super node V new is calculated according to the latitudes φ i and φ j of nodes v i and v j and the traffic weights f i，a and f j，b of their associated roads, so as to ensure that the position of the super node is closer to the node with larger traffic, reflecting the importance of the node in the traffic network. Similarly, the longitude λ new of the super node V new is calculated according to the longitudes λ i and λ j of nodes v i and v j .

[0118] inheriting all the attributes of nodes v i and v j to the super node V new , including the connected road edges, traffic signal information and historical data, etc. Specifically, the higher-level traffic signal configuration in the original node is retained to ensure that the merged network still meets the traffic management requirements. The super node V ne ​​w to all original associated nodes to maintain the connectivity and topology of the network. Merge node v i and the historical congestion records of v j , and establish a time sequence feature vector. And remove the edges e between the original connection nodes v i and v j from the network graph.

[0119] S104, generate candidate recommendation results containing candidate points on the backbone network, the candidate points including mandatory points, optional points and dynamic adjustment points.

[0120] It should be noted that the candidate recommendation results containing candidate points are generated on the backbone network, and the candidate points include mandatory points, optional points and dynamic adjustment points, so as to comprehensively consider the network structure, traffic conditions and passenger demand, generate diversified recommended points, and meet the needs of different passengers for carpooling, improve the flexibility and convenience of carpooling. At the same time, the introduction of dynamic adjustment points makes the recommended results adapt to the changes of traffic conditions in real time.

[0121] Further optionally, the step of generating candidate recommendation results containing candidate points on the backbone network, the candidate points including mandatory points, optional points and dynamic adjustment points, comprises:

[0122] Select all nodes with degree ≥ degree threshold as core pickup points, and define them as mandatory points;

[0123] For each pair of mandatory points, take the midpoint position of the shortest path as a candidate pickup point, and define it as an optional point;

[0124] According to the real-time folding state, generate temporary recommended points in the super node neighborhood as dynamic adjustment points.

[0125] Understandably, select all nodes with degree ≥ degree threshold (such as 3) as core pickup points, and define them as mandatory points. The node degree reflects the local properties of the node in the road network, that is, the number of edges directly connected to the node. Nodes with higher node degree have higher traffic flow and stronger connectivity, which are key locations for passengers to get on and off or vehicles to stop.

[0126] For each pair of mandatory points, take the midpoint position of the shortest path as a candidate pickup point, and define it as an optional point. Taking the midpoint position of the shortest path as a candidate pickup point can make the passenger pickup and drop-off position more reasonable, reduce the passenger walking distance or waiting time, and improve the convenience and comfort of carpooling. At the same time, the midpoint position is also convenient for vehicle parking and passenger distribution.

[0127] According to the real-time folding state, temporary recommended points are generated in the super node neighborhood as dynamic adjustment points. The real-time folding state reflects the real-time changes of the traffic conditions, which may make the original recommended points no longer suitable for the current traffic demand. And the super node is a key node in the traffic network, with high traffic flow and connectivity. The neighborhood is established around the super node, which can cover a wider range of passenger demand. The temporary recommended points generated in the super node neighborhood can be dynamically adjusted according to the real-time traffic conditions (such as passenger request coordinates, traffic flow, etc.), so that the recommended points are more in line with the passenger's boarding and alighting demand, thereby improving the success rate of carpooling. And the temporary recommended points as dynamic adjustment points can be adjusted in real time according to the current number of carpooling requests and resource conditions (such as the number of available vehicles). This helps to optimize resource allocation and avoid situations such as resource waste (such as empty vehicle running) or resource shortage (such as long waiting time for passengers).

[0128] Further optionally, the step of generating temporary recommended points in the super node neighborhood according to the real-time folding state as dynamic adjustment points comprises:

[0129] An initial circular neighborhood is established around the super node V new , and the service range of the adjacent super node V new is excluded in combination with the Voronoi diagram constraint.

[0130] The neighborhood radius is dynamically adjusted according to the current number of unmatched carpooling requests and the number of available vehicles, and the formula is:

[0131] ,

[0132] Where r dynamic is the dynamically adjusted neighborhood radius, r0 is the initially set reference neighborhood radius, k is the second adjustment coefficient, N requests is the current number of unmatched carpooling requests, and N vehicles is the number of available vehicles.

[0133] Based on the neighborhood, temporary recommended points are generated with multiple objectives.

[0134] The priority of each temporary recommended point is calculated by integrating the traffic matching degree, demand density and facility importance.

[0135] The candidate points are dynamically adjusted.

[0136] Understandably, an initial circular neighborhood is established around the super node V new , and the size of the initial circular neighborhood can be set according to actual conditions to cover a certain range around the super node.

[0137] In combination with the Voronoi diagram constraint, the service range of the adjacent super node V newThe service range of the Voronoi diagram can be used to determine the influence range of the traffic hub, thereby optimizing the distribution of traffic flow. Through the Voronoi diagram constraint, it can be avoided that the recommended points are too dense or overlapped, and it can be ensured that each recommended point can cover the effective passenger demand.

[0138] According to the number of current unmatched carpool requests and the number of available vehicles, the neighborhood radius is dynamically adjusted. Specifically, the neighborhood radius can be adjusted in real time by considering the proportion of the number of unmatched carpool requests and the number of available vehicles, so as to adapt to different traffic demands. When the number of unmatched carpool requests is large and the number of available vehicles is small, the neighborhood radius can be appropriately expanded to cover more potential passengers and improve the success rate of carpooling. Conversely, when the number of unmatched carpool requests is small and the number of available vehicles is large, the neighborhood radius can be appropriately reduced to avoid that the recommended points are too dispersed or redundant.

[0139] In generating the temporary recommended points, multiple targets such as traffic flow, passenger demand, facility importance, etc. can be considered comprehensively, so as to more comprehensively reflect the actual situation of the traffic network and improve the accuracy and effectiveness of the recommended points. In the neighborhood dynamically adjusted, the recommended points can be ensured to be located within the current reasonable range, so as to be more adaptive to the current traffic demand. Then the priority of each temporary recommended point can be calculated by comprehensively considering the flow matching degree, demand density and facility importance.

[0140] Based on the adjustment mechanism, the candidate points are dynamically adjusted. Specifically, the candidate point recalculation is periodically performed, and the points with high matching success rates in the historical candidate points are retained, so as to adapt to the changes of traffic demand and ensure the real-time and effectiveness of the recommended results. When road closure, severe congestion or special weather conditions are detected, local neighborhood reconstruction is immediately performed to regenerate candidate points that adapt to the current traffic conditions, thereby improving the success rate of carpooling. For the candidate points that have not been selected for a long time or have low matching success rates, the degradation or removal processing is performed. These candidate points may no longer be suitable as recommended points due to changes in traffic demand, facility adjustment, etc. Through the degradation or removal processing, the candidate point set can be optimized, and the accuracy and effectiveness of the recommended results can be improved.

[0141] Further optionally, the step of generating temporary recommended points based on the neighborhood and multiple targets comprises:

[0142] Sampling the flow velocity gradient along the neighborhood boundary road, and selecting a position with an absolute value of the flow velocity gradient exceeding a gradient threshold value as a temporary recommended point;

[0143] Applying a clustering algorithm to cluster the passenger request coordinates, and selecting a clustering core point as a temporary recommended point;

[0144] Identifying POIs, and selecting a POI closest to the super node V new as a temporary recommended point.

[0145] It can be understood that multiple targets such as traffic flow, passenger demand, facility importance, etc. can be comprehensively considered to generate temporary recommended points, so as to more comprehensively reflect the actual situation of the traffic network and improve the accuracy and effectiveness of the recommended points.

[0146] In a specific implementation, the flow rate gradient can be sampled along the neighborhood boundary road, and the position where the absolute value of the flow rate gradient exceeds the gradient threshold (such as 0.2 times the average flow rate) is selected as the temporary recommended point. The flow rate gradient reflects the rate of change of traffic flow, and the position where the absolute value exceeds the threshold represents an area where the traffic flow changes greatly, which is usually a hotspot position for passengers to get on or off or for vehicles to stop.

[0147] The clustering algorithm is applied to cluster the passenger request coordinates, and the clustering core point is selected as the temporary recommended point. In the traffic network, the clustering algorithm can be used to identify the dense areas of passenger request coordinates, and the passenger demand in these areas is high, so as to improve the success rate of carpooling by selecting the recommended point.

[0148] POIs (points of interest) such as bus stops and parking lot entrances are identified, and POIs close to the super node V_new are selected as temporary recommended points. These facilities are usually important places for passengers to get on or off, and as recommended points, they can facilitate passengers to get on or off and improve the convenience and comfort of carpooling.

[0149] S105, the candidate recommended result is optimized and screened to obtain the final recommended pickup point.

[0150] It should be noted that the candidate recommended result is optimized and screened to obtain the final recommended pickup point. The candidate points can be comprehensively evaluated and screened and optimized by considering factors such as walking distance cost, vehicle detour cost, and passenger matching probability. The optimized recommended pickup point not only considers the convenience of passengers, but also takes into account the driving efficiency of vehicles and the matching probability of passengers, thereby improving the overall efficiency of carpooling and passenger satisfaction.

[0151] Further, the step of optimizing and screening the candidate recommended result comprises:

[0152] A candidate pickup point set S = {mandatory point + optional point + dynamically adjusted point} is established;

[0153] For each carpooling request, the walking distance cost C walk , the vehicle detour cost C detour , and the passenger matching degree P match of each point in the candidate pickup point set S are calculated.

[0154] The comprehensive cost C is calculated by a comprehensive cost function, and the point with the minimum comprehensive cost C is selected as the recommended pickup point, and the formula is:

[0155] C = ω 1 C walk + ω 2 C detour - ω 3 P match Wherein, ω 1, ω 2, ω 3 are weights of walking distance cost, vehicle detour cost and passenger matching probability respectively.

[0156] It can be understood that the mandatory point, the optional point and the dynamic adjustment point are integrated to establish a candidate drop-off point set, and a comprehensive and flexible candidate drop-off point set is established by comprehensively considering the mandatory point, the optional point and the dynamic adjustment point, so as to cover a wider passenger demand and improve the success rate of carpooling.

[0157] And then, the walking distance cost C walk (the walking distance from the current location of the passenger sending a carpooling request to the candidate point), the vehicle detour cost C detour (the detour distance of the vehicle from the current location to the candidate point) and the passenger matching degree P match (the distance between the destination of the passenger sending a carpooling request and the destination of the passenger to be car-pooled on the vehicle, and the closer the actual distance, the higher the passenger matching degree) are comprehensively considered to calculate the comprehensive cost C of each candidate point in the candidate drop-off point set S, and the point with the minimum comprehensive cost C is selected as the final recommended drop-off point, so as to filter out the optimal drop-off point and improve the success rate of carpooling.

[0158] In summary, the carpooling drop-off point recommendation method for intercity network car-hailing provided by the application obtains intercity traffic network data, including road nodes, edges and weight information thereof, for representing the structure and connection relationship of the road network; the dynamic weight of each road is calculated based on real-time traffic data, the calculation of the dynamic weight can reflect the change of the traffic condition in real time, and the timeliness and accuracy of the recommended drop-off point are improved; the secondary roads are folded according to the dynamic weight and the folding threshold to generate a backbone network, which simplifies the network structure, and the generation of the backbone network not only improves the calculation efficiency, but also makes the selection of the recommended drop-off point more focused on the key positions, and improves the accuracy of the recommendation; the candidate recommended results containing each candidate point are generated on the backbone network, the candidate points include the mandatory point, the optional point and the dynamic adjustment point, so as to comprehensively consider the network structure, the traffic condition and the passenger demand, generate diversified recommended points, the diversified recommended points can meet the carpooling demand of different passengers, improve the flexibility and convenience of carpooling, and meanwhile, the introduction of the dynamic adjustment point makes the recommended results adapt to the change of the traffic condition in real time; the candidate recommended results are optimized and screened to obtain the final recommended drop-off point, and this step can comprehensively evaluate and screen the candidate points by considering the walking distance cost, the vehicle detour cost and the passenger matching probability, so that the final recommended drop-off point can not only consider the convenience of the passengers, but also take into account the driving efficiency of the vehicle and the matching probability of the passengers, thereby improving the overall efficiency of carpooling and the passenger satisfaction.

[0159] The application can reflect the change of traffic conditions in real time, generate more accurate recommended pickup points, improve the efficiency of carpooling, meet the carpooling needs of different passengers and improve the satisfaction of passengers by generating diversified candidate recommended points and optimizing the candidate recommended results, and the generation of the backbone network and the optimization of the candidate points effectively improve the calculation efficiency, so that the recommendation system can respond to passenger requests more quickly.

[0160] Embodiment two

[0161] Please refer to Figure 2 The application provides a carpool pickup point recommendation system for intercity network car hailing, which comprises:

[0162] The data acquisition module is used to acquire intercity traffic network data, including road nodes, edges and weight information.

[0163] The dynamic weight module is used to calculate the dynamic weight of each road based on real-time traffic data.

[0164] The folding module is used to fold the secondary roads according to the dynamic weight and the folding threshold to generate a backbone network.

[0165] The candidate generation module is used to generate a candidate recommended result containing candidate points on the backbone network, and the candidate points include mandatory points, optional points and dynamically adjusted points.

[0166] The recommended point generation module is used to optimize and screen the candidate recommended result to obtain the final recommended pickup point.

[0167] Further, the data acquisition module is also used to:

[0168] Acquire GIS road data of the target area, including node coordinates, road length and road grade.

[0169] An initial traffic network graph G=(V, E) is established, wherein V is a set of intersection nodes, and E is a set of road edges.

[0170] An initial weight w(e) is assigned to each edge e E. 初始 = road grade x basic speed coefficient.

[0171] Further, the dynamic weight module is also used to:

[0172] Access a real-time traffic data interface to acquire the current average speed v real (e) of each road segment.

[0173] The dynamic weight is calculated according to the formula:

[0174] w(e)=α‧v real(e) δ(t) + β · road grade · θ(e) + γ · accident probability coefficient · φ(w),

[0175] wherein δ(t) is a time penalty coefficient, θ(e) is a road function coefficient, and φ(w) is a weather influence coefficient.

[0176] Further optionally, the folding module is further configured to:

[0177] set an initial folding threshold value;

[0178] iterate through all edges e E;

[0179] dynamically calculate a folding threshold value θ = average dynamic weight x first adjustment coefficient;

[0180] if the dynamic weight w(e) < the folding threshold value θ, perform a folding operation to merge the nodes v i and v j at both ends of the edge e into a supernode V new ;

[0181] update the associated edge set to redirect the edges originally connected to the nodes v i and v j to the supernode V new , and recalculate the dynamic weights of the affected edges;

[0182] repeat the iteration until the network size converges.

[0183] Further optionally, the folding module is further configured to:

[0184] calculate the latitude coordinate of the supernode V new , according to the formula:

[0185] ,

[0186] wherein φ new is the latitude of the supernode, m is the number of roads associated with the node v i , φ i is the latitude coordinate of the node v i , f i，a is the traffic weight of the a-th associated road of the node v i , n is the number of roads associated with the node v j , φ j is the latitude coordinate of the node v j , and f j，b is the traffic weight of the b-th associated road of the node v j ;

[0187] calculate the longitude coordinate of the supernode V new , according to the formula:

[0188] ,

[0189] where λ new is the longitude of the supernode, λ i is the longitude coordinate of the node v i , and λ j is the longitude coordinate of the node v j ;

[0190] inherit all attributes of the nodes v i and v j to V new , including connected road edges, traffic signal information, and historical data;

[0191] remove the edge e originally connecting the nodes v i and v j from the network graph.

[0192] Further optionally, the candidate generation module is further configured to:

[0193] select all nodes with a degree ≥ degree threshold as core pickup points, as mandatory points;

[0194] for each pair of mandatory points, take the midpoint position of the shortest path therebetween as a candidate pickup point, as an optional point;

[0195] generate temporary recommended points in the supernode neighborhood according to the real-time folding state, as dynamic adjustment points.

[0196] Further optionally, the candidate generation module is further configured to:

[0197] establish an initial circular neighborhood centered on the supernode V new , and exclude the service range of the adjacent supernode V new in combination with Voronoi diagram constraints;

[0198] dynamically adjust the neighborhood radius according to the number of current unmatched carpool requests and the number of available vehicles, according to the formula:

[0199] ,

[0200] where r dynamic is the dynamically adjusted neighborhood radius, r0 is the initially set reference neighborhood radius, k is a second adjustment coefficient, N requests is the number of current unmatched carpool requests, and N vehicles is the number of available vehicles;

[0201] generate temporary recommended points based on the neighborhood, with multiple objectives;

[0202] Calculate the priority of each temporary recommended point by integrating the flow matching degree, demand density and facility importance;

[0203] Adjust the candidate points dynamically.

[0204] Further optionally, the candidate generating module is further configured to:

[0205] Sample the flow gradient along the neighborhood boundary road, and select the position with an absolute value of flow gradient exceeding a gradient threshold as a temporary recommended point;

[0206] Apply a clustering algorithm to cluster the passenger request coordinates, and select the clustering core point as a temporary recommended point;

[0207] Identify POIs, and select the POI closest to the super node V new as a temporary recommended point.

[0208] Further optionally, the recommended point generating module is configured to:

[0209] Establish a candidate pickup point set S = {mandatory point + optional point + dynamically adjusted point};

[0210] For each carpool request, calculate the walking distance cost C walk , vehicle detour cost C detour and passenger matching degree P match of each point in the candidate pickup point set S;

[0211] Calculate the comprehensive cost C by integrating the cost function, and select the point with the minimum comprehensive cost C as the recommended pickup point, the formula is:

[0212] C = ω1C walk + ω2C detour - ω3P match , wherein ω1, ω2, ω3 are the weights of the walking distance cost, vehicle detour cost and passenger matching probability, respectively.

[0213] The above-described embodiments only express several embodiments of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the present application. It should be noted that for ordinary skilled persons in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for recommending ride-sharing pick-up points for intercity ride-hailing services, characterized in that, The method includes: Acquire intercity transportation network data, including road nodes, edges, and weight information; Based on real-time traffic data, calculate the dynamic weight of each road; Secondary paths are folded according to dynamic weights and folding thresholds to generate the backbone network; Generate candidate recommendation results containing each candidate point on the backbone network. The candidate points include mandatory points, optional points, and dynamically adjusted points. The candidate recommendation results are optimized and filtered to obtain the final recommended boarding point; The step of calculating the dynamic weight of each road includes: Connect to the real-time traffic data interface to obtain the current average vehicle speed v of each road segment. real (e); The dynamic weights are calculated using the following formula: w(e) = α‧v real (e)‧δ(t)+β‧Road grade‧θ(e)+γ‧Accident probability coefficient‧φ(w), Where δ(t) is the time penalty coefficient, θ(e) is the road function coefficient, and φ(w) is the weather impact coefficient; The step of folding secondary paths according to dynamic weights and folding thresholds to generate the backbone network includes: Set the initial folding threshold; Traverse all edges e∈E; The dynamic calculation of the folding threshold θ = average dynamic weight × first adjustment coefficient; If the dynamic weight w(e) < the folding threshold θ, then a folding operation is performed to fold the nodes v at both ends of edge e. i and v j Merge into supernode V new ; Update the set of associated edges to include the original edge set connected to node v. i and v j Connected edges are redirected to supernode V new And recalculate the dynamic weights of the affected edges; Repeat the iteration until the network size converges.

2. The method for recommending ride-sharing pick-up points for intercity ride-hailing services according to claim 1, characterized in that, The steps for acquiring intercity transportation network data include: Obtain GIS road data for the target area, including node coordinates, road length, and road grade; Establish an initial traffic network graph G=(V,E), where V is the set of intersection nodes and E is the set of road edges; Assign an initial weight w(e) to each edge e∈E. 初始 = Road grade × Basic vehicle speed coefficient.

3. The method for recommending ride-sharing pick-up points for intercity ride-hailing services according to claim 1, characterized in that, The two nodes v at both ends of edge e i and v j Merge into supernode V new The steps include: Calculate the V of the supernode new The latitude coordinates are given by the formula: ， Where, φ new m is the dimension of the supernode, and m is the dimension of node v. i The number of associated roads, φ i For node v i latitude coordinates, f i，a For node v i The traffic flow weight of the a-th associated road, where n is the value of node v. j The number of associated roads, φ j For node v j latitude coordinates, f j，b For node v j The traffic flow weight of the associated road in Article b; Calculate the V of the supernode new The formula for longitude coordinates is: ， Where, λ new Let λ be the longitude of the supernode. i For node v i longitude coordinates, λ j For node v j Longitude coordinates; node v i and v j All properties are inherited by V new This includes the connected roadside, traffic signal information, and historical data; The original connection node v i and v j Edge e is removed from the network graph.

4. The method for recommending ride-sharing pick-up points for intercity ride-hailing services according to claim 1, characterized in that, The step of generating candidate recommendation results containing each candidate point on the backbone network, wherein the candidate points include mandatory points, optional points, and dynamically adjusted points, includes: Select all nodes with a degree greater than or equal to the degree threshold as core boarding points and designate them as mandatory points; For each pair of mandatory points, the midpoint of the shortest path is selected as a candidate boarding point and thus defined as an optional point. Based on the real-time folding status, temporary recommended points are generated within the neighborhood of the supernode as dynamic adjustment points.

5. The method for recommending ride-sharing pick-up points for intercity ride-hailing services according to claim 4, characterized in that, The step of generating temporary recommended points within the neighborhood of the supernode based on the real-time folding state, as dynamic adjustment points, includes: With supernode V new An initial circular neighborhood is established centered on the node V. Using Voronoi diagram constraints, adjacent supernodes V are excluded. new The scope of services; The neighborhood radius is dynamically adjusted based on the number of currently unmatched ride-sharing requests and the number of available vehicles, using the following formula: ， Where, r dynamic The neighborhood radius is dynamically adjusted, r0 is the initially set baseline neighborhood radius, k is the second adjustment coefficient, and N is the value of N. requests N represents the number of currently unmatched ride-sharing requests. vehicles Number of available vehicles; Temporary recommendation points are generated based on neighborhood and with multiple objectives. Priority is calculated for each temporary recommended point based on a combination of traffic flow matching degree, demand density, and facility importance; The candidate points are dynamically adjusted.

6. The method for recommending ride-sharing pick-up points for intercity ride-hailing services according to claim 5, characterized in that, The step of generating temporary recommendation points based on neighborhood and multiple objectives includes: Sample the velocity gradient along the neighborhood boundary road, and select the location where the absolute value of the velocity gradient exceeds the gradient threshold as a temporary recommended point; Clustering algorithms are applied to cluster the coordinates requested by passengers, and the core points of the clusters are selected as temporary recommended points. Identify POIs and select the distance from the supernode V. new The nearest POI is used as a temporary recommendation point.

7. The method for recommending ride-sharing pick-up points for intercity ride-hailing services according to claim 1, characterized in that, The steps for optimizing and filtering candidate recommendation results include: Establish a set of candidate boarding points S = {required points + optional points + dynamically adjusted points}; For each ride-sharing request, calculate the walking distance cost C to each point in the candidate pick-up point set S. walk Vehicle detour cost C detour passenger matching degree P match ; The comprehensive cost C is calculated using a comprehensive cost function, and the point with the minimum comprehensive cost C is selected as the recommended entry point. The formula is as follows: C=ω1C walk +ω2C detour -ω3P match ω1, ω2, and ω3 are the weights of walking distance cost, vehicle detour cost, and passenger matching probability, respectively.

8. A ride-sharing pick-up point recommendation system for intercity ride-hailing services, used to implement the ride-sharing pick-up point recommendation method for intercity ride-hailing services as described in any one of claims 1 to 7, characterized in that, The system includes: Data acquisition module: used to acquire intercity transportation network data, including road nodes, edges, and weight information; Dynamic weighting module: used to calculate the dynamic weight of each road based on real-time traffic data; Folding module: Used to fold secondary paths according to dynamic weights and folding thresholds to generate the backbone network; Candidate generation module: used to generate candidate recommendation results containing each candidate point on the backbone network. Candidate points include mandatory points, optional points, and dynamically adjusted points. Recommendation point generation module: used to optimize and filter candidate recommendation results to obtain the final recommended boarding point; The dynamic weight module is further used for: Connect to the real-time traffic data interface to obtain the current average vehicle speed v of each road segment. real (e); The dynamic weights are calculated using the following formula: w(e) = α‧v real (e)‧δ(t)+β‧Road grade‧θ(e)+γ‧Accident probability coefficient‧φ(w), Where δ(t) is the time penalty coefficient, θ(e) is the road function coefficient, and φ(w) is the weather impact coefficient; The folding module is also used for: Set the initial folding threshold; Traverse all edges e∈E; The dynamic calculation of the folding threshold θ = average dynamic weight × first adjustment coefficient; If the dynamic weight w(e) < the folding threshold θ, then a folding operation is performed to fold the nodes v at both ends of edge e. i and v j Merge into supernode V new ; Update the set of associated edges to include the original edge set connected to node v. i and v j Connected edges are redirected to supernode V new And recalculate the dynamic weights of the affected edges; Repeat the iteration until the network size converges.

Images